EP1131599B1 - Device for measuring the dimension and controlling of defects in optical fibres during production - Google Patents

Description

The present invention relates to a method for high speed and high accuracy absolute measurement of the diameter of transparent fibers and the control of internal and superficial defects.

• [1] Thèse de l'université de Saint-Etienne, par Lionel Delaunay, 28/03/1986
• [2] US Patent n° 3,982,816 de Laurence S. Watkins, Sept. 28, 1976
• [3] US Patent n° 4,027,977 de Ralph E. Frazee, June 7th, 1977 .
• [4] US Patent n° 4,280,827 de Edward F. Murphy, July, 28th, 1981 .
• [5] US Patent n° 4,176,961 de Ralph E. Frazee, December 4th, 1979 ,
• [6] EP 0 549 914 B1, de Button Leslie James , Corning Inc. 07/05/1997 et documents antérieurs.
• [7] 3396052966 FCOTAC, Alcatel FO

Since 1970, many studies have been devoted to the study of the propagation of light perpendicular to the axis of the fiber. Over the years, this work has led to numerous methods and solutions for dimensional measurement of fibers and control of defects. Among these works to which we have referred:

• [1] Thesis of the University of Saint-Etienne, by Lionel Delaunay, 28/03/1986
• [2] US Patent No. 3,982,816 to Laurence S. Watkins, Sept. 28, 1976
• [3] US Patent No. 4,027,977 to Ralph E. Frazee, June 7th, 1977 .
• [4] US Patent No. 4,280,827 to Edward F. Murphy, July, 28th, 1981 .
• [5] US Patent No. 4,176,961 to Ralph E. Frazee, December 4th, 1979 ,
• [6] EP 0 549 914 B1, by Button Leslie James , Corning Inc. 07/05/1997 and earlier documents.
• [7] 3396052966 FCOTAC, Alcatel FO

A method of measuring the diameter of an optical fiber is also described in the document US 5,513,004 .

The production of telecommunication optical fiber by fiber drawing, requires dispersions around the specified nominal diameter of increasingly weak and a lack of critical defect for the mechanical and optical characteristics of the fibers but especially for their lifetime. This is the case of defects in the glass, "air-line" or in the coating, "bubbles", "delaminations", as well as the eccentricity of the glass in its polymer coating. Finally, the increase in fiberizing speeds and the high frequency fluctuations of the fiber diameter require faster instruments than those commonly used to visualize these phenomena and reduce them.

The present invention relates to a method and an assembly for both dimensional measurement and control of glass fibers before coating (bare fiber) as dimensional measurement and control of coatings (coating). In a more general way it applies to any rod of circular section, optically transparent, corresponding to the figure 1 and having an index profile R1, N1 and R2, N2. No known device to date can achieve contactless, more than 50 KHz, and statically an absolute diameter measurement with resolutions of the order of a few hundredths of a micron including the detection of defects.

References [1] and [2] summarize the fundamental principles of light propagation in a transparent fiber, as described figure 1 , in particular when it is illuminated by a monochromatic field of light colimated, coherent, perpendicular to the axis of the fiber. The deviation of the light by the fiber, produces a system of interferometry fringes whose periods and phases depend on the profiles of indices; diameters, materials, homogeneity, concentricity. From the analysis of these fringes, according to the angles of measurement, one deduces precise information on the conformity of the fiber with respect to a model defined as absolute reference. The quality of the measured signal represented by the contrast and the regularity of the fringes as well as by the energy of the signal, depends on the optical quality of the fiber, ie the absence of defects, the homogeneity, the geometric regularity . Continuous contrast analysis of high-speed fringes can detect very small defects such as bubbles in polymer coatings.

The reference [1] describes in detail the propagation relations and deduces a method for rapidly measuring the diameter variations and a method for measuring the diameter by calibration on a standard fiber. The cited documents describe assemblies for measuring diameter variations by counting fringe slip or diameter by measuring the period of the fringes with reference to a period of standard fiber. The document [6] uses a Fourier transform processing method for measuring the diameter and detecting small defects in the glass.

The figure 1 recalls the propagation phenomena in a transparent cylindrical structure consisting of a center of radius "r" and index N ₁ , and an envelope of radius "R" and index N ₂ , whose ratio r / R is, at most of the order of 0.5, usual case of telecommunication fibers. This structure represents in particular either a multi-mode fiberglass with its core (62.5 / 125μm) or a monomode fiber, or finally a glass fiber coated with its coating (125 / 245μm). We consider that the indices of the two coatings of coating compared to the glass to consider the same model. However, the invention can be applied with models taking into account the differences in the indices of the coating layers.

avec

ψi

Relation de propagation de chaque raie laser contribuant à la constitution du signal au même point.

n

indice N1 de la matière

θ

Angle de mesure

D

diamètre extérieur

λ

Longueur d'onde du laser

The propagation of light rays through the fiber has been widely studied and we will only use the results of this work. The reflected (Rr) and transmitted (Rt) beams interfere infinitely around the fiber. The model that describes the contribution of the beams to the fringe system is expressed by: $$\mathrm{S}=\underset{\mathrm{i}}{\Sigma }{\mathrm{\psi }}_{\mathrm{i}}\left(\mathrm{not},\mathrm{\theta },\mathrm{D},\mathrm{\lambda }\right),$$

with

ψi

Propagation relationship of each laser line contributing to the constitution of the signal at the same point.

not

N1 index of the material

θ

Angle of measurement

D

outside diameter

λ

Laser wavelength

In an angular zone θ between 40 ° and 80 °, two functions Ψ corresponding to the rays Rt and Rr, largely dominate the others. We collect the contribution of the other residual beams (multiple reflections in the fiber) under the character ε which leads to reduce then the relation (1) to: $$\mathrm{S}={\mathrm{\Psi }}_{\mathrm{rt}}+{\mathrm{\Psi }}_{\mathrm{rr}}+\mathrm{\epsilon }$$

avec $$\mathrm{f}\left(\mathrm{n},\mathrm{\theta }\right)=\sin \left(\mathrm{\theta }/\mathrm{2}\right)+{\left[{\mathrm{n}}^{\mathrm{2}}+\mathrm{1}-\mathrm{2}*\mathrm{n}*\cos \left(\mathrm{\theta }/\mathrm{2}\right)\right]}^{\mathrm{1}/\mathrm{2}}$$

soit en nombre de franges : $$\mathrm{N}=\mathrm{\Delta }/\mathrm{\lambda }=\left(\mathrm{D}/\mathrm{\lambda }\right)*\mathrm{f}\left(\mathrm{n},\mathrm{\theta }\right)+\mathrm{¼}$$

$$\mathrm{N}\left(\mathrm{0}\right)=\left(\mathrm{D}/\mathrm{\lambda }\right)*\left({\mathrm{n}}^{\mathrm{2}}-\mathrm{2}*\mathrm{n}+\mathrm{1}\right)+\mathrm{¼}$$

The difference Δ of the two beams Ψ is expressed as a function of the outer diameter of the fiber "D", the outer coating index "n = N ₁ ", the wavelength "λ" and the angle of measurement "θ", according to references [1] and [2]. $$\mathrm{\Delta }=\mathrm{D}*\mathrm{f}\left(\mathrm{not},\mathrm{\theta }\right)+\mathrm{\lambda }/\mathrm{4}$$

with $$\mathrm{f}\left(\mathrm{not},\mathrm{\theta }\right)=\sin \left(\mathrm{\theta }/\mathrm{2}\right)+{\left[{\mathrm{not}}^{\mathrm{2}}+\mathrm{1}-\mathrm{2}*\mathrm{not}*\cos \left(\mathrm{\theta }/\mathrm{2}\right)\right]}^{\mathrm{1}/\mathrm{2}}$$

in number of fringes: $$\mathrm{NOT}=\mathrm{\Delta }/\mathrm{\lambda }=\left(\mathrm{D}/\mathrm{\lambda }\right)*\mathrm{f}\left(\mathrm{not},\mathrm{\theta }\right)+\mathrm{¼}$$

$$\mathrm{NOT}\left(\mathrm{0}\right)=\left(\mathrm{D}/\mathrm{\lambda }\right)*\left({\mathrm{not}}^{\mathrm{2}}-\mathrm{2}*\mathrm{not}+\mathrm{1}\right)+\mathrm{¼}$$

If we know precisely each parameter, λ, θ and n, the relations (3) and (4) constitute a reference model for the absolute measurement of D, without calibration constraint by a standard fiber. λ, is a precisely measurable or known parameter of the laser (He). "N1" is the known index precisely of the outer layer of the fiber. θ, must be measured on the instrument. Each elementary point of the sensor 6 ( figure (2 ) used must therefore correspond to a precise angular value.

The originality of the presented method rests, on the one hand, on the digital processing of the signal which identifies a parameterizable model on the signal and which leads directly to the absolute measurement of the diameter, on the other hand, an analog method of rapid measurement very high resolution diameter variations, and finally, the correlation of the results of the numerical and analog methods to provide a fast and absolute measurement method. A fundamental element of the method lies in the optical quality of the signal, characterized by the contrast of the fringes. The permanent analogue control of the contrast, included in the device, also leads to the detection of possible defects of the fiber.

Absolute Diameter Measurement

This measurement is carried out by a digital processing of a signal coming from a sensor 6 ( figure 2 ), constituted in our case of a CCD strip, placed between 40 and 80 ° with respect to the origin of the laser beam. In the experimental setup, the angular window of the bar is 24 °. Digital processing is characterized by a low frequency of measurements, currently of the order of 10 Hz.

et $$\mathrm{S}=\mathrm{M}+\mathrm{\epsilon }$$

A representation model of the fringe system produced by Rt and Rr can be written: $$\mathrm{M}=\left[\mathrm{AT}/\left(\mathrm{B}+\mathrm{\theta }\right)\right]*\left[\mathrm{1}+{\mathrm{<figure-callout\; id="2"\; label="Sin"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">Sin</figure-callout>}}^{\mathrm{2}}\left\{\mathrm{NOT}\left(\mathrm{D},\mathrm{\theta }\right)*\mathrm{\pi }+\mathrm{\phi }\right\}\right],$$

and $$\mathrm{S}=\mathrm{M}+\mathrm{\epsilon }$$

The identification of the model M closest to the fringe system S leads to determining A, B, D and φ minimizing the residual ε. All methods of identification are usable. For our part we use a least squares minimization method, on the one hand on the period by optimizations of D and φ, then on the amplitude. Parameters A and B allow the search for an approximate amplitude profile allowing the detection of defects. Their values are approximately proportional to the diameter (energy of the signal) which makes it possible to initialize them after adjustment of the apparatus. Only the value of D, in N (D, θ), interests us for the measurement. φ is a parameter which makes it possible to better adjust the phase of the model to that of the signal.

The angular calibration of the sensor is done by reference to the direction of the axis of incidence of the measuring laser taken as angular value 0. To identify the angular positions of the pixels, the mounting of the figure 10 . The optical bench of the measuring apparatus is mounted on a test bench consisting on the one hand of an optical slot (14) to limit the field and the energy of the laser on the sensors, of a large angular encoder precision, centered on the axis of the fiber (axis of the apparatus), of an integral assembly, consisting of an optical (12) focusing of the laser beam and a sensor (13) giving a voltage proportional to the position of the laser spot. The assembly (12) (13) is rotated angularly until the sensor (13) indicates the position of the laser beam at 0 volts. The angular encoder is then secured to the assembly (12) (13) and initialized to zero on the laser axis. The assembly (12) (13) is then rotated by any angle, of the order of 110.degree. For example to enable angular positioning and referencing of a mirror on the axis of the fiber. The mirror and the assembly (12) (13) are adjusted to find an indication of 0 volts on the sensor (13). The mirror is then made integral with the encoder, the angular value θ ₀ corresponding to 0 volts of the encoder is read and the assembly (12) (13) is disconnected. The mirror is then referenced angularly. The axis of the laser deflected by the mirror will then be A = 2 * (π - θ) - θ ₀ . By turning the mirror (11) the laser spot is displaced on the sensors 6 and 7. It is thus possible to point out point by point the angular positions of the pixels and to identify the transfer function angle = f (pixels) including the optical system 4.

The fringe system in an angular zone of 24 ° taken between 40 and 80 °, for a fiber D = 125 μm, λ = 780 nm and n = 1.478 is shown figure 3 . The precise measurement of the evolution of the period of the signal measured as a function of the angle will make it possible to identify D in the model.

A first approach is done by identifying the parameters D and φ, by minimizing the differences between the model and the measured signal, between θ1 and θ2 which correspond to the angular range of the CCD array. In this approach, we do not take into account the phase of the fringe system but only the period as a function of the angle.

• Indice : de la silice connue avec une incertitude négligeable en fonction de la longueur d'onde. Pour les coating, les paramètres permettant de calculer les indices sont fournis avec précision par les fabriquants.
• Longueur d'onde du laser : Pour un laser gaz, les valeurs sont connues aussi avec grande précision et incertitude négligeable. Pour une diode laser stabilisée, la longueur d'onde est mesurée à ± 0,2 nm sur 780, soit ± 0,025%.
• Valeurs angulaires : La résolution du capteur est de 3 secondes d'arc avec une incertitude totale cumulée de la calibration de l'ordre de 1 minute d'arc, soit sur 70° (4200 minutes d'arc) 0,024.

We obtain a stability and reproducibility of measurements better than 0,02μm. The total uncertainty calculated on D is less than ± 0.15 μm. These uncertainties are as follows:

• Index: known silica with negligible uncertainty as a function of wavelength. For coatings, the parameters for calculating the indices are accurately provided by the manufacturers.
• Laser wavelength: For a gas laser, the values are also known with great accuracy and negligible uncertainty. For a stabilized laser diode, the wavelength is measured at ± 0.2 nm over 780 ± 0.025%.
• Angular values: The resolution of the sensor is 3 arc seconds with a cumulative total uncertainty of the calibration of the order of 1 minute of arc, ie over 70 ° (4200 arc minutes) 0.024.

The total uncertainty of the apparatus would therefore be 0.05% of the diameter, ie on 125 μm, ± 0.06 μm. To these uncertainties must be added fiber imperfections (ovality, local optical disturbances, residual signals "ε" which reduce the total uncertainty to an estimated value of ± 0.15μm.

The angular measurement range between 58 and 82 ° corresponds to about 30 fringes. Relation (6) shows that θ = 0, N (0) varies proportionally to the Diameter, that is to say that the phase of Rt, hence from S to θ = 0 varies proportionally to the diameter. If the model is safe enough to determine the phase of the wave (Rt, Fig. 1 ) passing through the fiber for θ = 0, then it is possible to measure with a very high precision the diameter of the fiber. In this case, the "φ" phase identified in the first approach takes on its full meaning. It is then possible to envisage, starting from the angular range of measurement of the CCD, to raise doubts on the number of fringes of the model, between θ = 0 and a fringe on the CCD. In fact, the total uncertainty of the measurements made in an angular window of 24 ° is 0.15%. The number of fringes between 0 and 70 ° is about 180 to λ = 780 nm. The uncertainty on the number of fringes between 0 and 70 ° amounts to 0.15% * 180 = 0.28, which is considerably less than half a fringe. The phase of S at the origin of the angles is then added to the model to obtain a much improved measurement accuracy. The function (4) is taken into account in the angular range of 0 to θ _o on the CCD.

The advantage of this method of measurement by adaptation of the model comes from the fact that it frees itself from the fluctuations related to the variations of the residual modulations of the signal in the angular range of measurement. It takes into account the model of N in the whole of the measurement angle between 0 ° and θ. This gives the digital measurement a very high stability, of the same order as the analog measurement that we discuss below.

However, there remain uncertainties related to the geometric dispersion of the fiber compared to the model that led to the establishment of basic relationships: circularity of the fiber, cylindricity defects, as well as defects related to the divergence of the laser beam. Although these defects are very limited to diameter, they do not however allow to achieve an absolute measuring device in accuracies corresponding to the resolutions reached. On the other hand, the best measurement methods that could allow us to validate the metrological results are no better than ± 0.15 μm. This means that, in this configuration, the first evaluation, by difference of angles, is sufficient.

Rapid measurements of Diameter variations

With respect to a given angular position, the sliding of a fringe represents the result of the variation of all the fringe periods between the angular origin and the angular position of the measurement plus the phase variation at the angular origin. This phase measurement method at the point θ has the advantage of being very stable, very reproducible, very sensitive to variations in the diameter and not very sensitive to the residual modulations of the fringe system. It differs from the digital measurement method but leads to the same resolutions. This method gives only the variations of the diameter but at high speed, in opposition to the absolute method, numerical, but slow. They are therefore complementary to ensure both the absolute measurement and the high speed.

The variation of the Diameter leading to this slip is expressed: $$\mathrm{dD}=-\mathrm{\lambda }*\mathrm{dN}/\mathrm{f}\left(\mathrm{not},\mathrm{\theta }\right)$$

The pitch of the fringes as a function of the angle θ is shown Figure 9 . It can be seen that the pitch of the fringes remains substantially constant in the zone 60 ° ± 10 °. This characteristic makes it possible to place in this zone sensors spaced at a constant pitch corresponding to the pitch of the fringes. It is also interesting for the measurement of the period (frequency) by Fourier transform [6].

From the relation (9) we can extract a rapid measurement of the Diameter variations at high speed and high resolution. For example, for an angle of 70 °, a wavelength of 780 nm and a glass index of 1.478, the slip of dN = ¼ fringe represents a variation of the diameter of 0.13 μm. In our method, we easily obtain a resolution of 1/32 fringe, or 0.01μm.

We find in the literature assemblies for the measurement of Diameter variations. They rely on counting methods of binarized signals S1, S2 whose minimum resolution remains% fringe. The difference of our method is to perform a binarization of the fringes signal in a different way that will allow to count the slip in increment of% fringe also but also to interpolate this slip within an increment.

• La mesure ne prend pas en compte les modulations résiduelles d'amplitude du signal ni les fluctuations des périodes mais seulement la phase des franges par rapport aux capteurs. Cela rend la mesure insensible aux modulations de l'amplitude des franges.
• Le glissement de la phase des franges résulte de la variation de toutes les franges depuis l'origine angulaire (axe du laser) ce qui fourni une très grande sensibilité. On obtient une résolution et une reproductibilité de l'ordre de 0,01µm.
• Cette mesure permet la mise en oeuvre de méthodes analogiques continues avec des bandes passantes élevées, 75 Khz dans notre cas.

This measurement is made by placing sensors at a pitch corresponding to a proportion of the nominal pitch of the fringes, see Figure 7 . The interest of this approach is multiple:

• The measurement does not take into account the residual modulations of signal amplitude nor the fluctuations of the periods but only the phase of the fringes with respect to the sensors. This makes the measurement insensitive to modulations of the amplitude of the fringes.
• The sliding of the fringes phase results from the variation of all the fringes since the angular origin (laser axis) which provides a very high sensitivity. A resolution and a reproducibility of the order of 0.01 μm are obtained.
• This measurement allows the implementation of continuous analog methods with high bandwidths, 75 Khz in our case.

• S1 = C1-C3
• S2 = C2-C4

pour éliminer les composantes continues.The Figures 6 and 7 present the principle of this high-resolution measure. As already known, a succession of optical sensors (pixels) spaced from an odd number of ¼ of the optical signal fringe is placed. For example, the sensors C1 to C4 can be used ( Figure 7 ) which are spaced ¼ of fringe or C'1 to C'4 which are spaced de of fringe. This last arrangement makes it possible to use sensors with pitch between pixels that are larger than the pitch of the fringes to be measured. The combination of the sensors C1 to C4 is reproduced over several periods and they are connected in parallel so as to add the signals of Ci of the same index between them. The sensitive surfaces of the sensors must not exceed ¼ of the pitch of the fringes to obtain a jump contrast in the signals. We choose 1/5 in our case with a satisfactory result. The following analog sum of the signals is then produced:

• S1 = C1-C3
• S2 = C2-C4

to eliminate the continuous components.

• A1 = f[S1-S2>0]
• A2 = f[S1+S2<0]

Instead of binarizing S1> 0 and S2> 0, as commonly practiced to obtain counting / down counting signals, the following conditions are binarized:

• A1 = f [S1-S2> 0]
• A2 = f [S1 + S2 <0]

1. 1) /A1*/A2 : s1/s2 = S2/S1
2. 2) A1*/A2 : s1/s2 = -S1/S2
3. 3) A1*A2 : s1/s2 = -S2/-S1
4. 4) /A1*A2 : s1/s2 = S1/-S2

The two binary signals A1 and A2 will serve two functions, one of counting and counting of the sliding of S1 and S2 by ¼ of a period, in a conventional way, the other to switch the signals S1, -S1, S2 and - S2 on a circuit of calculation of the arctgente (s1 / s2) according to the four combinations between A1 and A2 marked 1 to 4 in the Figure 7 (/ Ai logical complement of Ai).

1. 1) / A1 * / A2: s1 / s2 = S2 / S1
2. 2) A1 * / A2: s1 / s2 = -S1 / S2
3. 3) A1 * A2: s1 / s2 = -S2 / -S1
4. 4) / A1 * A2: s1 / s2 = S1 / -S2

The figure 6 shows a block diagram of the realization of an analog measurement on this principle from signals S1 and S2. The signals A1 and A2 drive a dual multiplexer which alternately switches the four inputs S1, -S1, S2, -S2 to s1 and s2 as indicated above. Analogue measurement tests provided the ENR1 record in terms of reproducibility.

At initialization, digital and analog measurement methods are combined to efficiently filter the dispersions between the two methods. This is interesting in production where signal fluctuations between two measurements can be considered as random. For this we compare the values measured in digital and analog, sampled at the same time, by averaging the "n" last measurements. The residual dispersions of the digital measurement in the first phase of the treatment (diameter calculated between two angular values) are then weighted and the analogical absolute value is initialized according to the difference of the averages and not on the difference of the instantaneous measurements.

Defect detection

The defects are essentially small air tubes, "air-line", in the glass fibers, or localized defects, superficial.

In coatings (coating), they include air bubbles, inclusions of unmelted materials, delaminations (gas gap between the glass and the coating), thicknesses, under thicknesses and eccentricity.

In this setup, we limit ourselves to real-time detection of optical disturbances produced by air-lines, bubbles, delaminations and inclusions. The variations of diameters are taken into account simply by comparisons of the fast analog output with programmed or regulated thresholds.

Two parameters are used to characterize the quality of the optical signal: the energy, which increases with the size of an air-line, but especially the contrast of the fringes (difference between the maxima and the minima of S1 and S2), which constitutes a fundamental element for this method of measurement. The absence of contrast on the fringes effectively prevents the measurement.

The real-time energy is obtained by the sum of the currents coming from the photodiodes of the analog sensor. The real-time contrast is obtained by taking the maximum of the four signals S1, -S1, S2, -S2. This signal is naturally fluctuating since the Si are sinusoids.

The appearance of a defect, even small, produces either a sharp drop in contrast or saturation. Whatever the type of defect, the contrast always reacts much more sensitive than the energy. It is this parameter that we retain for the detection of defects.

For the finest defects of the air-line type, the digital processing extracts the component of the measured signal which does not correspond to an acceptable distribution of the amplitude and the period of the fringes (see figure 3 (normal signal), figure 4 (disturbed signal), figure 5 , difference of the signals of the Figures 3 and 4 ).

Claims (5)

1. A method for measuring the diameter of an optical fiber and for detecting defects in this fiber, comprising the steps consisting in:

   - directing a laser beam perpendicularly to the axis of the fiber, to create a system of interference fringes;

   - providing a first sensor (6) capable of receiving the fringe system and a first digital processing board (8) capable of analyzing the signal emitted by the first sensor (6);

   characterized in that the method further comprises the steps consisting in :

   - providing a second sensor (7) capable of receiving the fringe system and a second analogue processing board (9) capable of analyzing the signal emitted by the second sensor (7);

   - angularly calibrating the first and second sensors (6, 7);

   - achieving an absolute and periodic measurement of the diameter (D) of the fiber by determining, by means of the first sensor (6) and of the first processing board (8), parameters A, B, D, ϕ allowing the optimal identification of the fringe system and of the following physical model: $$M=\frac{A}{B+\theta }\times \left[1+{\sin }^2\left\{N\left(D,\theta \right)\times \pi +\varphi \right\}\right]$$

   

   $$N\left(D,\theta \right)=\frac{D}{\lambda }\times \left[\mathit{sin}\frac{\mathit{\theta }}{\mathit{2}}+{\left({\mathit{n}}^{\mathit{2}}+\mathit{1}-\mathit{2}\mathit{n}\times \mathit{cos}\frac{\mathit{\theta }}{\mathit{2}}\right)}^{\mathit{1}/\mathit{2}}\right]+\frac{1}{4}$$

   

   
   where M is the amplitude of the signal measured by the first sensor (6), N is the number of fringes, λ is the wavelength of the laser and θ is the angle of measurement;

   - in parallel, achieving a relative and continuous measurement of the variations of the diameter (D) of the fiber by analysis, by means of the second sensor (7) and of the second processing board (9), of the shift of the fringes, according to the following relationship: $$\mathit{dD}=-\lambda \times \frac{\mathit{dN}}{\mathit{sin}\frac{\theta }{2}+{\left(n^2+1-2n\times \mathit{cos}\frac{\theta }{2}\right)}^{1/2}}$$

   

   
   said second processing board (9), which is analogue, being initialized by said first processing board (8), which is digital, so that the method allows obtaining a quick, continuous, and absolute measurement of the diameter (D) of the fiber.
2. The method according to claim 1, characterized in that an optical system (4) is provided, projecting a portion of the laser beams deviated by the optical fiber, at an angular range between 40 and 80°, on the two sensors (6, 7), said sensors (6, 7) being optically aligned and located in the focal plane of the optical system (4).
3. The method according to claim 1 or 2, characterized in that the step of angular calibration of the first and second sensors (6, 7) is carried out using a calibration bench external to the sensor, allowing to accurately identify the axis of the laser beam in order to reference the angular encoder, and to angularly deviate the laser by a mirror integral to the angular encoder, on several points of the sensors, through the optical system, in order to accurately identify the angular positions of these points.
4. The method according to any of claims 1 to 3, characterized in that it further comprises the steps consisting in comparing two analogue signals (S1, S2) of the second sensor (7) to generate two logic signals (A, B) allowing, on the one hand, the digital counting of the shift of the fringes by quarter period, on the other hand, interpolating between each switching of A or B the continuous variation of the phase of the fringes by carrying out the arctangent of (S1 / S2), finally adding the two results - counting plus interpolation - to afford a quick, continuous and high-resolution measurement, in a wide variation range of the diameter.
5. The method according to any of claims 1 to 4, characterized in that it further comprises the steps consisting in using the loss of the fringes contrast to detect, in real time, by analogue means, the presence of defects in the fiber and by the use of the amplitudes difference of the model and of the real signal to detect, by digital processing, the very small defects.

Images